Abstract
Most of HCI (Human Computer Interfaces) applications are based on a graphical interface, mouse and keyboard; in this work we investigated the use of voice activation as input device to control the mouse pointer on the screen. The control of the windows icon mouse pointer (WIMP) by voice command is currently based on using vowel utterances, this category of letters is easy to learn and to be pronounced, especially for individuals who are physically disabled or have a partial voice disorder. In addition, vowels are quite easy to recognize by automatic speech recognition (ASR) systems. In this work we represent a system for the control of mouse cursor based on voice command, using the pronunciation of certain phonemes and short words. The Mel Frequency Cepstral Coefficients (MFCCs) and Predictive Coding (LPC) are selected as distinctive features. The non linear sequence alignment known as Dynamic Time Warping (DTW), Hidden Markov Models (HMMs) have been have been tested as classifiers for matching components (vowels and short words).